The topic of this thesis is about the investigation of methods in use for object recognition in facility management by using the example of office and administration buildings. Thereby, 360° spherical panoramas are used as data base. It is further evaluated to which extent DIN standards, workplace guidelines and laws can provide information to be used in the course of object classification. Description of both the technical basics for taking the images and the process for linking these images to a 360° spherical panorama will be addressed. Furthermore, an overview of the object recognition process with its individual sub-processes is also given. These sub-processes are image preparation, image pre-processing, image segmentation and feature extraction. To describe the classification process an overview of classic classification methods is presented using rules and methods based on machine learning technologies.

Investigating the application of a rule-based system for object classification in the context of office buildings would be beneficial. The aim is to enable classification of objects found in image analysis using rules that can be created from object knowledge. This is because systems using machine learning methods require large amounts of training data. This amount of data is not necessary in rule-based systems because existing object knowledge can be used to create rules.